Level Extreme platform
Subscription
Corporate profile
Products & Services
Support
Legal
Français
VFP CRC Function
Message
From
21/12/2012 18:04:02
 
 
To
21/12/2012 17:04:46
General information
Forum:
Visual FoxPro
Category:
Coding, syntax & commands
Environment versions
Visual FoxPro:
VFP 9 SP2
Miscellaneous
Thread ID:
01560390
Message ID:
01560439
Views:
47
Must guess I draw a blank on your description as well -
somehow a (third party ?) timestamp seems more adequate to my
current visualization, meaning I am not getting it ;-)

>No, I do not mean that - I am not using this function to create unique hashes of inputs. Without going into detaiI, I am using this function in a simple comms scenario, a handshake between 2 apps. The collision issue is not relevant for my use in this case.
>
>>By "the collision issue is not relevant", do you mean that the unreliability introduced is acceptably low?
>>
>>>Sorry Al but the exact details of my use in this case do matter. I am not using this function for cryptographic hashing or as a security function, for which I use proper hash algorithms as you suggest. I am using SYS(2007) for a different purpose and the collision issue is not relevant.
>>>
>>>UPDATE: For crypto requirements I use Chilkat Crypto library which are extensive in their coverage and work very well with VFP - http://www.chilkatsoft.com/crypt-activex.asp
>>>
>>>
>>>>The exact details of your use for it don't really matter. Ultimately, when hash functions are used, the purpose is to compare the resulting hashes to determine "same" or "different". With a high rate of collisions, CRC32 gives too many false positives i.e. reports "same" when really the underlying expressions are different. CRC32 is unreliable.
>>>>
>>>>It's a bit ironic, because people usually start investigating hash functions because they've taken an interest in reliability or security. In those circumstances, using a known unreliable function, when better ones are readily available, is counterproductive.
>>>>
>>>>>Its being used as a simple handshake check between 2 VFP applications. One of these applications is being rewritten in a different langauge and I was just looking at whether it would be easier to recode this function in the new language or whether to recode the SYS(2007) in the remaining VFP app to something else. Hence the collision thing is not an issue for the purpose to which it is being used.
>>>>>
Previous
Reply
Map
View

Click here to load this message in the networking platform